---
title: "MTCARS"
author: "Stefani Rmalho"
output: 
  flexdashboard::flex_dashboard:
    orientation: rows
    vertical_layout: fill
    theme: cerulean 
    source_code: embed
    social: menu
---

```{r setup, include=FALSE}
### Fonte de pesquisa ##
# Documentacao do flexdashboard
# https://rmarkdown.rstudio.com/flexdashboard/index.html

### referencias ###
# https://r4ds.had.co.nz/
# Garrett Grolemund
# Hadley Wickham

# Carregando os modulos
library(ggplot2)
library(dplyr)
library(tibble)
library(ggcorrplot) # Plotar as correlacoes
library(DT) # Imprimir o dataset permitindo interacoes
library(psych) # Funcao describe
library(knitr) # Imprimir tabelas em um formato mais amigavel
library(modelr) 
library(flexdashboard)

# Criando um dataset os dados mtcars
df <- datasets::mtcars %>%
  rownames_to_column() %>%
  as_tibble() %>%
  mutate(vs = factor(vs),
         am = factor(am),
         gear = factor(gear),
         carb = factor(carb),
         cyl = factor(cyl))

# Tema dos graficos
theme_dahs <- theme_light() +
  theme(panel.grid.minor.x = element_blank(),
  panel.grid.major.x = element_blank(),
  panel.grid.minor.y = element_blank())

# Criando os modelos de regressao
model_lm_1 <- lm(mpg ~ wt + disp,df)
model_lm_2 <- lm(mpg ~ wt * disp,df)
```

Resultados {data-icon="fas fa-chart-line"}
===================================== 

Row {data-height=20}
-----------------------------------------------------------------------
### Modelo A: mpg ~ wt + disp

### Modelo B: mpg ~ wt * disp

Row {data-height=130}
-----------------------------------------------------------------------

### R2

```{r}
# calculando o R2 do modelo 1
r2_1 <- round(rsquare(model_lm_1,df) * 100,2)

valueBox(r2_1, 
         icon = "fas fa-percent",
         color = "primary")
```

### RMSE

```{r}
# calculando o RMSE do modelo 1
rmse_1 <- round(rmse(model_lm_1,df),4)

valueBox(rmse_1, 
         icon = "fas fa-chart-area",
         color = "primary")
```

### R2

```{r}
# calculando o R2 do modelo 2
r2_2 <- round(rsquare(model_lm_2,df) * 100,2)
valueBox(r2_2, 
         icon = "fas fa-percent",
         color = "primary")

```

### RMSE

```{r}
# calculando o RMSE do modelo 2
rmse_2 <- round(rmse(model_lm_2,df),4)

valueBox(rmse_2, 
         icon = "fas fa-chart-area",
         color = "primary")

```

Row {data-height=500}
-----------------------------------------------------------------------

### Valores previstos x reais

```{r}
# Hitograma valores previstos x mpg
p5 <- df %>% 
  add_predictions(model_lm_1) %>% 
  ggplot() +
  geom_histogram(aes(x = mpg, bins = 10), size = 1, fill = "sky blue") +
  geom_histogram(aes(x = pred, bins = 10), size = 1, fill = "orange") +
  theme_dahs

p5
```

### Frequência dos residuos

```{r}
# grafico com a frequencia dos residuos
p6 <- df %>% 
  add_residuals(model_lm_1) %>% 
  ggplot(aes(x = resid)) +
  geom_freqpoly(bins = 40, color = "blue", size = 1) +
  theme_dahs

p6
```

### Valores previstos x reais

```{r}
# Hitograma valores previstos x mpg
p7 <- df %>% 
  add_predictions(model_lm_2) %>% 
  ggplot() +
  geom_histogram(aes(x = mpg, bins = 10), size = 1, fill = "sky blue") +
  geom_histogram(aes(x = pred, bins = 10), size = 1, fill = "orange") +
  theme_dahs

p7
```

### Frequência dos residuos

```{r}
# grafico com a frequencia dos residuos
p8 <-  df %>% 
  add_residuals(model_lm_2) %>% 
  ggplot(aes(x = resid)) +
  geom_freqpoly(bins = 40, color = "blue", size = 1) +
  theme_dahs

p8
```

Row {data-height=350}
-----------------------------------------------------------------------

###

```{r}
# resudo do modelo
p9 <- summary(model_lm_1)[c(1,4)]
p9
```
###

```{r}
# resumo do modelo
p10 <- summary(model_lm_2)[c(1,4)]
p10
```

Analise Exploratória {data-icon="fa-signal"}
===================================== 

Row {data-height=500}
-----------------------------------------------------------------------

### Distribuição de frequência por MPG

```{r}
# Plot p0 - Histograma do atributo mpg
p0 <- ggplot(df, aes(x = mpg)) +
  geom_histogram(bins = 10, fill = "SKY blue") +
  theme_dahs

p0
```

### Boxplot do atributo MPG por CYL

```{r}
# Plot p1 - Box plot mpg x cyl
p1 <- ggplot(df, aes(x = cyl, y = mpg)) +
  geom_boxplot(fill = "orange", color = "orange3") +
  theme_dahs

p1
```

### Relação entre o atributo MPG x WT

```{r}
# Plot p1 - scatterplot mpg x wt
p2 <- ggplot(df, aes(x = mpg, y = wt)) +
  geom_point(size = 4, color = "red") +
  geom_point(size = 3, color = "orange") +
  geom_smooth(method = lm, se = FALSE, color = "grey2") +
  theme_dahs

p2
```

### Análise de Correlação

```{r}
# Plot p3 - correlacoes
p3 <- df %>% select_if(is.numeric) %>%
  cor() %>%
  ggcorrplot(hc.order = TRUE,
             type = "lower",
             outline.color = "white")

p3
```

Row {data-height=500}
-----------------------------------------------------------------------


### Estatística Descritiva {.no-mobile}

```{r}
# plot p4 - tabela estatistica
p4 <- df %>%
  describe() %>%
  data.frame() %>%
  rownames_to_column() %>%
  rename(Atributo = rowname) %>%
  select(-vars, -n) %>%
  slice(-1) %>%
  kable()

p4
```



Dataset {data-icon="fa-table"}
=====================================   

### Dataset MTCARS {.no-mobile}
    
```{r}
# Tabela
t <- datatable(df)
t